Date: Sun, 14 Nov 2010 11:16:18 +0200 From: Mikolaj Golub <to.my.trociny@gmail.com> To: freebsd-fs@freebsd.org Cc: Pawel Jakub Dawidek <pjd@freebsd.org> Subject: hastd/primary.c micro fixes Message-ID: <86wrogqn8d.fsf@kopusha.home.net>
next in thread | raw e-mail | index | archive | help
[-- Attachment #1 --]
noticed the following in hastd/primary.c:
1) in init_remote() when incoming connection is set up, if proto_client()
fails it only reports about the error and try proto_connect() anyway. This
will cause a crash as proto_conn structure is not valid. Primary should rather
exit (as it is when outgoing connection is set up, and as it is in the patch
below) or set proto_conn pointer to NULL and goto close.
2) in guard_thread() timeout.tv_sec is set in loop, while it looks like it may
be set only once before the loop.
--
Mikolaj Golub
[-- Attachment #2 --]
Index: sbin/hastd/primary.c
===================================================================
--- sbin/hastd/primary.c (revision 215165)
+++ sbin/hastd/primary.c (working copy)
@@ -577,7 +577,7 @@ init_remote(struct hast_resource *res, struct prot
* Setup incoming connection with remote node.
*/
if (proto_client(res->hr_remoteaddr, &in) < 0) {
- pjdlog_errno(LOG_WARNING, "Unable to create connection to %s",
+ primary_exit(EX_TEMPFAIL, "Unable to create connection to %s",
res->hr_remoteaddr);
}
/* Try to connect, but accept failure. */
@@ -2008,6 +2008,7 @@ guard_thread(void *arg)
PJDLOG_VERIFY(sigaddset(&mask, SIGINT) == 0);
PJDLOG_VERIFY(sigaddset(&mask, SIGTERM) == 0);
+ timeout.tv_sec = RETRY_SLEEP;
timeout.tv_nsec = 0;
signo = -1;
@@ -2033,7 +2034,6 @@ guard_thread(void *arg)
guard_one(res, ii);
lastcheck = now;
}
- timeout.tv_sec = RETRY_SLEEP;
signo = sigtimedwait(&mask, NULL, &timeout);
}
/* NOTREACHED */
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?86wrogqn8d.fsf>